Computational framework and machine learning approach to fractional order soil helminth infections disease model for control mechanism

Soil-transmitted helminth (STH) infections are exceedingly common, estimated to afflict 1.5 billion individuals worldwide, or 24% of the total population. The aim of this work is to model the transmission dynamics of soil-transmitted helminth (STH) infections using fractional-order calculus integrated with machine learning to enhance prediction accuracy and control strategies. Specific fundamental characteristics like the non-negativity of solutions and the invariant region within where the model equations hold epidemiological significance. Equilibrium points and basic reproductive number are explored. Based on the basic reproduction number, the stability of the model's equilibrium points is investigated. When the fundamental reproduction number is smaller than one, the disease-free equilibrium is demonstrated to be asymptotically stable both locally and globally, suggesting the possibility of curing these infections. Sensitivity analysis identifies key parameters. The analysis shows that basic reproduction number is highly sensitive to transmission rate, recruitment rate, and parasite carrying capacity. The existence and uniqueness of the fractional model's solutions are proven using a fixed-point method. Machine learning techniques, such as Artificial Neural Networks (ANN) with Levenberg-Marquardt backpropagation (ANNLMB), Random Forest (RF), and Support Vector Machines (SVM) are employed to solve the model and analyze transmission patterns. The ANNLMB approach achieved low mean square errors, validating model. Moreover, RF and SVM classifiers attained prediction accuracies of 99.83% and 99.5%, respectively, effectively distinguishing between infection classes. Integrated control measures and machine learning-enhanced models can support real-time monitoring and targeted interventions, improving public health and life expectancy.